The language learning theories of Professor J. Cummins

This page has information about the second language learning theories of J. Cummins and their implications for mainstream teachers.


Professor J. Cummins is one of the world's leading authorities on bilingual education and second language acquisition. This page covers four of his most influential insights: different types of language (BICS/ CALP), common underlying language proficiency, a way to assess task difficulty, and additive/subtractive bilingualism.

@ Mainstream teachers who have a knowledge of Cummins' insights will be in a much stronger position to help the ESL students in their classes learn the mainstream subject while at the same time developing their English language proficiency. ~


Cummins makes the distinction between two differing kinds of language proficiency. BICS are Basic Interpersonal Communication Skills. These are the "surface" skills of listening and speaking which are typically acquired quickly by many students. This is particularly true for students from language backgrounds similar to English who spend a lot of their school time interacting with native speakers.

CALP is Cognitive Academic Language Proficiency, and, as the name suggests, is the basis for a student's ability to cope with the academic demands placed upon her in the various subjects.

Cummins states that while many children develop native speaker fluency (i.e. BICS) within two years of immersion in the target language, @@ it takes between 5-7 years for a child to be working on a par with native speakers as far as academic language is concerned. ~~

In other words, the non-native speakers in mainstream classes who have exited from the ESL program Non-native English students typically remain in special English language (ESL) programs for a maximum of three years. This is clearly not enough time to reach the levels CALP attained by their native English-speaking peers. are still, in most cases, in the process of catching up with their native speaking peers as far as CALP is concerned.

- Implications for mainstream teachers

It should not be assumed that non-native students who have attained a high degree of fluency and accuracy in everyday spoken English (BICS) have the corresponding academic language proficiency (CALP).

Students may well still be struggling with the demands of comprehending and composing academic texts. In particular, they are very likely to still have considerable gaps in their academic vocabulary.

Common Underlying Proficiency

Briefly stated, Cummins believes that in the course of learning one language a child acquires a set of skills and implicit metalinguistic knowledge that can be drawn upon when working in another language. This common underlying proficiency (CUP), as he calls these skills and knowledge, is illustrated in the Iceberg Analogy below.

It can be seen that the CUP provides the base for the development of both the first language (L1) and the second language (L2). It follows that any expansion of CUP that takes place in one language will have a beneficial effect on the other language(s). This theory also serves to explain why it becomes easier to learn each additional language.

CUP (Common Underlying Proficiency)
Wikimedia Commons

- Implications for mainstream teachers

@@@ It is very important that students be encouraged to continue their native language development. When parents ask about the best ways they can help their child at home, you can reply that the child should have the opportunity to read extensively in her own language. ~~~

You could suggest that parents make some time every evening to discuss with their child, in their native language, what she has done in school that day. For example, ask her to talk about the science experiment she did, question her about her understanding of primary and secondary sources of historical information, have her explain how she has solved a mathematics problem.

As Cummins (2000) states:

"Conceptual knowledge developed in one language helps to make input in the other language comprehensible."

If a child already understands the concepts of justice or honesty in her own language, all she has to do is acquire the label for these terms in English. She has a far more difficult task, however, if she has to acquire both the label and the concept in her second language.

Assessing task difficulty

Cummins has devised a model whereby the different tasks we expect our students to engage in can be categorized. In the diagram below tasks range in difficulty along one continuum from cognitively undemanding to cognitively demanding. And along the other continuum from context-embedded to context-reduced.

A context-embedded task is one in which the student has access to a range of additional visual and oral cues. For example, he can look at illustrations of what is being talked about or ask questions to confirm understanding.

A context-reduced task is one such as listening to a lecture or reading dense text, where there are no sources of help other than the language itself.

Clearly, a D quadrant task, which is both cognitively demanding and context- reduced, is likely to be the most difficult for students. This is particularly true for non-native speakers in their first years of learning English, i.e. before they have acquired much CALP. However, it is essential that ESL students develop the ability to accomplish such tasks, since academic success is impossible without it.

Click inside a quadrant above to see examples of that type.

For more on assessing task difficulty, using Bloom's taxonomy (Wikipedia), click the button below.

Assessing task difficulty

Bloom's taxonomy (Knowledge → Comprehension → Application → Analysis → Synthesis) provides a useful way of determining whether a task is demanding or undemanding. Activities which fall within the category of Knowledge - such as collecting, naming, showing etc. - will clearly be less demanding than Analysis activities such as comparing, explaining and inferring.

The degree to which a task is context-embedded depends on the number of channels of information available to the student. So a student who listens to a news report on the radio has only one channel of information - this is a context-reduced learning experience.

Compare this with the student who reads a report about the same event in a newspaper article which contains photographs and diagrams. The student can read at her own speed and has access to a dictionary.

If she can also ask another student or her parents to explain parts of the text, then she has many channels of information available to her. This is clearly a context-embedded activity and as a result is much more manageable.

It is difficult to see the value of any tasks that are cognitively undemanding and context-reduced. Copying a list of the kings and queens of England from a textbook to an exercise book is an example of such an activity. It is sometimes called busy work.

- Implications for mainstream teachers

If teachers have an awareness of the likely difficulty of a task, based on Cummins' model, they can judge its appropriateness for the non-native speakers in their classes and in this way avoid much frustration.

This does not mean, however, that ESL students should be fed a diet of cognitively-undemanding tasks. It may be beneficial to use such activities in the student's early days at school, in order to build confidence, or as a lead in to a more challenging activity.

However, teachers should switch soon to tasks that engage the students' thinking processes, and make these tasks accessible by providing visual or other support. Once students are comfortable with such activities, they can be gradually exposed to tasks that are both cognitively-demanding and context-reduced.

For an interesting discussion of what happens when teachers start with a D quadrant task and then have to modify it to avoid embarrassment and confusion in the classroom, see the Mackay article listed in the references below.

Coelho, see reference below, has a useful example of how to use the quadrant in designing support for ESL students doing a science project.

Additive/subtractive bilingualism

Cummins draws the distinction between two types of bilingualism. Additive bilingualism is bilingualism in which the first language continues to be developed and the first culture to be valued while the second language is added.

Subtractive bilingualism is bilingualism in which the second language is added at the expense of the first language and culture, which diminish as a consequence.

Cummins (1994) quotes research which suggests students working in an additive bilingual environment succeed to a greater extent than those whose first language and culture are not valued by their schools and by the wider society.

- Implications for mainstream teachers

The dangers of subtractive bilingualism for the non-native English speakers in international schools are obviously not so strong as, say, for the children of immigrants to the USA.

Nevertheless, we should do all we can to demonstrate to non-native English students that their cultures and languages are equally as valid and valued as the Anglo/American culture and English language that inevitably dominates English-medium international school life.

Teachers and departments should explore ways to incorporate the different cultural backgrounds of our students into their daily teaching and curricula.


Cummins' research in the areas of second language acquisition discussed above date from the last century. But his insights and advice continue to have great relevance for language policy makers, administrators and mainstream teachers alike.


The summary above is based on the following articles or book extracts by Cummins or about his theories.

  • ^ 1 ^
  • ^ 63 ^
  • ^ 10 ^
  • ^ 13 ^
  • ^ 17 ^
  • ^ 14 ^
  • ^ 16 ^
  • ^ 15 ^
  • ^ 22 ^
  • ^ 31 ^
  • ^ 43 ^